// llil2grt.cpp.
// Inspired by llilgrt.pl: improve sort performance via a negative count.
// g++ compile on Linux:
//    g++ -o llil2grt -std=c++11 -Wall -O3 llil2grt.cpp
// This g++ command also works with mingw C++ compiler (https://sourceforge.net/projects/mingw-w64)
// that comes bundled with Strawberry Perl (C:\Strawberry\c\bin\g++.exe).
// Example run: llil2grt tt1.txt tt2.txt tt3.txt >out.txt

#include <cstddef>
#include <cstdint>
#include <cstdlib>
#include <cstring>
#include <ctime>
#include <cstdio>

#include <string>
#include <vector>
#include <set>
#include <map>

#include <algorithm>
#include <utility>

#include <iostream>
#include <fstream>
#include <sstream>

static_assert(sizeof(size_t) == sizeof(int64_t), "size_t too small, need a 64-bit compile");

// ----------------------------------------------------------------------------

// Crude hack to see Windows Private Bytes in Task Manager by sleeping at
// program end (see also sleep hack at end of main)
//    #include <chrono>
//    #include <thread>

// ----------------------------------------------------------------------------

typedef long long llil_int_type;
using str_int_type     = std::pair<std::string, llil_int_type>;
using int_str_type     = std::pair<llil_int_type, std::string>;
using map_str_int_type = std::map<std::string, llil_int_type>;
using vec_str_int_type = std::vector<str_int_type>;
using set_int_str_type = std::set<int_str_type>;

// Mimic the Perl get_properties subroutine ----------------------------

// Limit line length and use ANSI C functions to try to boost performance
#define MAX_LINE_LEN_L 255

static void get_properties(
   int                nfiles,      //  in: the number of input files
   char*              fname[],     //  in: the input file names
   map_str_int_type&  hash_ret)    // out: a hash of properties
{
   FILE* fh;
   char line[MAX_LINE_LEN_L+1];
   char* word;
   llil_int_type count;

   for (int i = 0; i < nfiles; ++i) {
      fh = ::fopen(fname[i], "r");
      if (fh == NULL) {
         std::cerr << "Error opening '" << fname[i] << "' : errno=" << errno << "\n";
         continue;
      }
      while ( ::fgets(line, MAX_LINE_LEN_L, fh) != NULL ) {
         word  = ::strtok(line, "\t");
         count = ::atoll( ::strtok(NULL, "\n") );
         hash_ret[word] -= count;
      }
      ::fclose(fh);
   }
}

// ---------------------------------------------------------------------

int main(int argc, char* argv[])
{
   if (argc < 2) {
      std::cerr << "usage: llil2grt file1 file2 ... >out.txt\n";
      return 1;
   }

   std::cerr << "llil2grt start\n";
   time_t  tstart1 = ::time(NULL);
   clock_t cstart1 = ::clock();

   // Create the hash of properties
   map_str_int_type hash_ret;
   get_properties(argc - 1, &argv[1], hash_ret);

   clock_t cend1  = ::clock();
   double ctaken1 = (double) (cend1 - cstart1) / (double)CLOCKS_PER_SEC;
   std::cerr << "get_properties      CPU time : " << ctaken1 << " secs\n";

   clock_t cstart2 = ::clock();

   // To avoid calling sort(), try creating an inverted std::set container
   // Note: negative count gives desired ordering (just like Perl GRT sort :)
   set_int_str_type s;
   for ( auto const& kv : hash_ret ) {
      // Note: emplace_hint() was a lot faster than emplace()
      s.emplace_hint( s.end(), kv.second, kv.first );
   }

   clock_t cend2s = ::clock();

   // Output the (already sorted) std::set - no sort() function required!
   // Note: fix up negative count via -n.first
   for ( auto const& n : s ) { std::cout << n.second << '\t' << -n.first << '\n'; }

   clock_t cend2 = ::clock();
   time_t  tend2 = ::time(NULL);

   long    ttaken   = static_cast<long>(::difftime(tend2, tstart1) + 0.5);
   double  ctaken   = (double) (cend2 - cstart1) / (double)CLOCKS_PER_SEC;

   double  ctaken2s = (double) (cend2s - cstart2) / (double)CLOCKS_PER_SEC;
   double  ctaken2o = (double) (cend2  - cend2s)  / (double)CLOCKS_PER_SEC;

   std::cerr << "emplace set sort    CPU time : " << ctaken2s << " secs\n";
   std::cerr << "write stdout        CPU time : " << ctaken2o << " secs\n";

   std::cerr << "total               CPU time : " << ctaken   << " secs\n";
   std::cerr << "total        wall clock time : " << ttaken   << " secs\n";

   // Hack to see Private Bytes in Windows Task Manager (uncomment next line so process doesn't exit too quickly)
   //   std::this_thread::sleep_for(std::chrono::milliseconds(90000000));

   return 0;
}